Message boards : Number crunching : ELAPSED TIME V. CPU TIME
Message board moderation
Author | Message |
---|---|
Send message Joined: 31 Dec 07 Posts: 1152 Credit: 22,363,583 RAC: 5,022 |
I was looking at the Famous graphics and I noticed the difference between the time displayed in the Boinc manager and that displayed in the model graphics. With the model at 67.76% the elapsed time (shown on the manager) is 183:17. The CPU time (shown on the model graphic interface) is only 154:19. THAT’S NEARLY 29 HOURS DIFFERENCE. At this rate the difference will be over 42 hours by the time the model finishes. That’s almost 2 full days. I always new that there was a difference between the “elapsed time†shown on the manager and the CPU time, but, I did not realize it was so great. Was the difference that great in the other types of model? We will never know for the CM’s as the Boinc Manager in use back then only displayed CPU time not elapsed time. Was the difference as great for the SM and AM3P’s? |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
Was the difference as great for the SM and AM3P’s? Potentially. Elapsed time measures the time when an application is scheduled by BOINC. CPU time measures the actual CPU usage. That means elapsed time includes any time the application is scheduled but pre-empted because the OS is running a higher priority (non-BOINC) application. My laptop has a HadSM3 task at 91% with 409 hours CPU time and 438½ hours elapsed time. A few months ago I was soak testing some of my work on my C2Q system for a week. I was running a HadAM3P task at the time and its elapsed time was almost double its CPU time. "The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer |
©2024 cpdn.org